Introduction

Column

Difference between classical music and movie soundtracks

My Corpus

Music has always been an integral part of human culture, evolving over time to encompass a wide range of genres and styles. With the advent of digital streaming platforms like Spotify, it has become easier than ever before to access and explore different types of music. Among the diverse range of music available on Spotify, two popular playlists are the “Classical Essentials” and “Iconic Soundtracks” playlists.

Classical music has a rich history spanning centuries and is characterized by its emphasis on formal structure, harmony, and melody. While classical music is often associated with traditional concert settings and academic study, it can also evoke a wide range of moods and emotions, from serene and contemplative to triumphant and dramatic. In contrast, soundtracks are typically composed to accompany visual media, such as movies or TV shows, and aim to evoke specific moods or emotions. Soundtracks may draw from a variety of musical genres, including classical music, as a source of inspiration or to create a particular atmosphere. However, it’s important to note that not all soundtracks are classical music, and not all classical music is used in soundtracks.

This portfolio will examine the extent to which classical music and soundtrack music are related and how they may have influenced each other. By comparing the musical features of these two genres, we hope to gain insights into the ways in which classical music has influenced soundtrack music. We aim to explore the similarities and differences between these genres and to identify the unique contributions of each to the musical landscape. Ultimately, our research will provide a better understanding of the relationship between classical music and soundtrack music.

For my portfolio, I have chosen a corpus that compares classical music with film soundtracks. The corpus will include track from the “Classical Essential” and “Iconic Soundtracks” playlists on Spotify. I am interested in this because both classical music and film soundtracks can evoke emotions in listeners, but the audiences of both can be very different, which is why I want to see what differentiates the two.

In my corpus I will compare classical compositions with film soundtracks. Within these groups, there are various natural subgroups such as individual composers, specific movies, and musical genres. I expect to find differences in instrumentation and production between the two groups, as well as differences between individual composers and genres. However, I also anticipate finding commonalities in melody, harmony, and musical structure.

The tracks in my corpus are representative of the groups I want to compare, but there are probably some recordings missing or variations that could affect the analysis. For example, there may be differences between the live and studio versions of a classical piece that are not reflected in the corpus. I will keep this in mind when the data.

One atypical track in my corpus is “Hedwig’s Theme” by John Williams, which for most of the track uses only the celesta and a unique melody. A typical track is “Für Elise” by Ludwig van Beethoven, which is a well-known classical piece with a recognizable melody.

Visualizations

What to expect

Column

Classical Essentials

Column

Iconic Soundtracks

Emotional features

Exploring Emotional Tone in Playlists: A Scatterplot Analysis of Energy, Loudness, and Acousticness


Visualization

The scatterplot displays a comparison of the two playlists based on their energy and loudness levels, with the color scale representing acousticness. The x-axis shows energy, which is a measure of the intensity and activity level of the music, while the y-axis shows loudness, which is a measure of the volume. The acousticness value represents the degree to which the track is acoustic (as opposed to electronic).

Why this plot

Energy measures the intensity and activity level of a song. High energy levels can create a feeling of excitement, while low energy levels can create a feeling of calm or relaxation. Loudness measures the volume of a song. High loudness levels can create a feeling of intensity or aggressiveness, while low loudness levels can create a feeling of intimacy or introspection. Acousticness measures the degree to which a song is acoustic (as opposed to electronic). High acousticness levels can create a feeling of nostalgia or emotional depth, while low acousticness levels can create a feeling of modernity or innovation.

Emotion within the plot

The scatterplot shows that the Classical Essentials playlist tends to have music with lower energy and loudness levels, which appear as a cluster of points in the lower left-hand corner of the plot. These songs also tend to have higher acousticness levels, as indicated by the warm colors on the color scale. This suggests a more relaxed and subdued emotional tone. This type of music may create a soothing atmosphere, which can be helpful for stress relief or relaxation. In contrast, the Iconic Soundtracks playlist has a wider range of energy and loudness levels, which appear as a more scattered distribution of points across the plot. This suggests a more dynamic and varied emotional tone. The high-energy tracks may create a feeling of excitement, while the low-energy tracks may create a feeling of melancholy or introspection.

The overlap between the two playlists, with some songs having similar energy and loudness levels, may suggest that there are emotional similarities between them. For example, some of the higher-energy tracks in the Iconic Soundtracks playlist may evoke similar emotions as the Classical Essentials playlist’s lower-energy tracks, such as feelings of relaxation or introspection. Similarly, some of the lower-energy tracks in the Iconic Soundtracks playlist may evoke similar emotions as the Classical Essentials playlist’s higher-energy tracks, such as feelings of excitement or inspiration. These emotional similarities may be related to the acousticness of the tracks or other factors, and may be subjective to the individual listener’s preferences and experiences. However, the scatterplot provides insight into the general emotional tone and atmosphere of the two playlists based on their energy, loudness, and acousticness levels.

In summary, the plot shows the differences and similarities between two playlists in terms of energy, loudness, and acousticness, which may reflect their overall emotional tone and atmosphere.

Tempo and Mode Revealing Emotional Differences


Visualization

This plot visualizes the relationship between the tempo and mode of the tracks within the two playlists.” The x-axis represents the mean tempo (beats per minute) of each track, while the y-axis represents the standard deviation of the tempo for each track. The size of the markers represents the time signature of each track, and the color represents the mode (major or minor) of each track. The opacity of the markers represents the loudness of each track.

Why this plot

The tempo and mode of a piece of music can convey emotional information to the listener. A faster tempo can create a feeling of excitement or urgency, while a slower tempo can create a feeling of calm or melancholy. The mode of a piece of music (major or minor) can also affect the emotional response of the listener, with major keys generally associated with happiness and minor keys associated with sadness.

Emotion within the plot

The plot shows that the tracks in the “Classical Essentials” playlist tend to have a lower mean tempo compared to the tracks in the “Iconic Soundtracks” playlist. This means that the tracks in the “Classical Essentials” playlist are generally slower in tempo than the tracks in the “Iconic Soundtracks” playlist. The lower tempo of the tracks in the “Classical Essentials” playlist can create a calming effect on the listener, as slower tempos are often associated with relaxation and meditation. On the other hand, the higher tempo of the tracks in the “Iconic Soundtracks” playlist can create a more energetic and dynamic emotional response in the listener.

Moreover, the tempo standard deviation for the “Classical Essentials” playlist is lower than that of the “Iconic Soundtracks” playlist. This suggests that the tracks in the “Classical Essentials” playlist have a more consistent tempo throughout the song, while the tempo in the “Iconic Soundtracks” playlist may vary more, creating a more dynamic emotional response in the listener.

Additionally, the plot shows that most of the tracks in the “Classical Essentials” playlist are in a major key, while the tracks in the “Iconic Soundtracks” playlist are evenly split between major and minor keys. Music in a major key is often associated with positive emotions such as happiness, joy, and optimism, while music in a minor key is often associated with negative emotions such as sadness, grief, and melancholy. Therefore, the fact that the “Classical Essentials” playlist has more tracks in a major key suggests that it may create a more uplifting emotional response in the listener compared to the “Iconic Soundtracks” playlist, which has a more varied key signature.

In summary, the plot suggests that the “Classical Essentials” and “Iconic Soundtracks” playlists may evoke different emotional responses from listeners due to differences in tempo and mode. The slower tempo and more consistent tempo of the “Classical Essentials” playlist, combined with its emphasis on major key signatures, suggest that it may create a more calming and uplifting emotional response in the listener. The more varied tempo and mode of the “Iconic Soundtracks” playlist, on the other hand, may create a wider range of emotional responses, including both positive and negative emotions.

Exploring the Emotional Content of Music Playlists: Valence-Energy Scatter Plot Analysis


Visualization

The graph depicts the relationship between energy and valence of songs in the two different playlists. The x-axis represents the valence, which is a measure of the musical positivity conveyed by a song, while the y-axis represents the energy, which is a measure of the intensity and activity of a song.

Why this plot

The emotional quadrants in the plot are based on two measures: valence and energy. Valence refers to the degree of positivity or negativity of an emotion, with high valence indicating positive emotions and low valence indicating negative emotions. Energy, on the other hand, refers to the level of intensity or activity in the music, with high energy indicating fast and loud music and low energy indicating slow and quiet music. The emotions chosen for the quadrants - angry, happy, sad, and calm - are commonly used in music psychology research and are based on the valence and arousal dimensions of emotion. Angry and happy are high valence and high arousal emotions, while sad and calm are low valence and low arousal emotions. The placement of songs in each emotional quadrant is influenced by their valence and energy levels. Songs with high energy and high valence are likely to be placed in the happy quadrant, while those with low energy and high valence are likely to be placed in the calm quadrant. Similarly, songs with low energy and low valence are likely to be placed in the sad quadrant, while those with high energy and low valence are likely to be placed in the angry quadrant.

Emotion within the plot

In this plot, both playlists are predominantly placed in the sad quadrant, which suggests that the majority of songs in these playlists have negative emotions with low energy levels. This is interesting, as one might expect the “Soundtracks” playlist to have more positive emotions, given the uplifting nature of many movie soundtracks. However, the plot also shows that there are some outliers in each playlist, which indicates that there are some songs with higher valence and energy levels that could potentially be used to create a more diverse emotional experience within the playlists.

Conclusion

Based on the earlier graphs provided, it may have been unexpected to see both playlists predominantly placed in the sad emotional quadrant. The tempo and loudness of the “Iconic Soundtracks” playlist seemed higher and therefore expected to be “happier” and “angrier”. Based on valence and energy alone this doesn’t seem the case and both turn out to be predominantly in the sad quadrent. Overall, this plot provides a useful visualization of the emotional content of songs in different playlists and highlights the importance of valence and energy in shaping emotional responses to music.

Chordograms

Chordograms Soundtracks


Visualization

Chromagrams are commonly used to visualize the pitch content of a piece of music over time.A chromagram is a visual representation of the pitch content of an audio signal, where the pitch information is projected onto a time-frequency plane. In a chromagram, the y-axis represents the pitch (in terms of the pitch class) and the x-axis represents time. The intensity of the color indicates the strength of that pitch class in the audio signal. In the context of music analysis, the chromagram is often used to study the harmonic content of a musical piece. By examining the chromagram, we can identify the chords that are being played at each point in time. This information can be used to analyze the structure of the music and to identify patterns and changes in the harmonic content.

In terms of emotion, the chromagram can give us an idea of the tonality of the music. Major keys tend to have a brighter, more uplifting sound, while minor keys tend to be more melancholy and introspective. Additionally, certain chord progressions are associated with specific emotions.

Chosen Songs

The first chromagram being discussed is an outlier in a playlist, which was positioned inside the “happy” quadrant of the scatterplot earlier.The second chromagram is a representative song from the playlist, which has energy and valence values that are the same as the average values for the playlist.

Emotion Within

Comparing the two chromagrams in this context, the outlier chromagram appears to have more varied and complex harmonic content, with more prominent use of minor chords and a greater overall range of pitches. This could suggest a more complex emotional state, with a greater range of emotions being expressed. The representative chromagram, on the other hand, appears to have a more stable and predictable harmonic structure, with a relatively even distribution of major and minor chords and a narrower range of pitches. This could suggest a more stable emotional state, with fewer shifts in mood or feeling. However, it’s important to note that these are just broad generalizations and that the emotional impact of a musical piece can vary greatly depending on a number of factors, including context, culture, personal experience, and individual perception.

Chordograms Classicals


Chosen Songs

The first chromagram represents an outlier song in the playlist, which was positioned in the “calm” quadruple instead of “sad”. The second represents a representative song that is closer to the mean energy and valence of the playlist.

Emotion Within

Comparing the two chromagrams in this context, the outlier chromagram appears to have more varied and complex harmonic content, with more prominent use of minor chords and a greater overall range of pitches. This could suggest a more complex emotional state, with a greater range of emotions being expressed. The representative chromagram, on the other hand, appears to have a more stable and predictable harmonic structure, with a relatively even distribution of major and minor chords and a narrower range of pitches. This could suggest a more stable emotional state, with fewer shifts in mood or feeling. However, it’s important to note that these are just broad generalizations and that the emotional impact of a musical piece can vary greatly depending on a number of factors, including context, culture, personal experience, and individual perception.

Visual Analysis

Notable patters in Iconic Soundtracks found by hierarchical clustering


Here a heatmap and dendrogram based on the audio features of 20 songs from the “Classical Essentials” playlist are visualized. The heatmap displays the audio features of 20 songs from the Iconic Soundtracks playlist, where each row represents a song, and each column represents an audio feature. The colors in the heatmap correspond to the values of each audio feature, where yellow indicates higher values and blue indicates lower values.

The dendrogram illustrates the results of hierarchical clustering of the songs based on their audio features. The height of the dendrogram branches indicates the degree of similarity between songs.

Notable patterns found in the dendrogram show that the 20 songs can be clustered into two main groups. The first group consists of seven songs that are closely related to each other, forming a cluster on the right side of the dendrogram. The second group consists of 13 songs, which are further divided into two subclusters on the left side of the dendrogram.

Among the audio features, loudness seems to be the most useful feature for clustering, as songs with similar loudness values are clustered together. On the other hand, features such as valence and speechiness appear to be less useful for clustering, as there are no clear patterns of similarity or dissimilarity based on these features.

Notable patters in Classical Essentials found by hierarchical clustering


Here a heatmap and dendrogram based on the audio features of 20 songs from the “Classical Essentials” playlist are visualized. The heatmap displays a color-coded grid where rows represent the songs and columns represent the audio features, such as danceability, energy, and tempo. Warmer colors indicate higher values while cooler colors represent lower values. The dendrogram is a tree-like diagram that shows the hierarchical relationships among the songs based on their audio feature similarities.

Looking at the dendrogram, we can observe that the songs form several distinct clusters, with some clusters being more closely related to each other than others. One notable pattern is that songs with similar tempos tend to cluster together. Additionally, songs with similar pitches also tend to group together. In contrast, the features related to the timbre of the music do not seem to have a strong effect on the clustering.

Regarding the heatmap, we can see that some features, such as loudness, energy, and danceability, show a relatively high degree of variability across the songs, while others, such as speechiness and instrumentalness, show little variability. Interestingly, the heatmap also shows that some songs have a distinctive profile of high or low values across multiple features, indicating that some songs share specific musical characteristics.

Overall, the dendrogram and heatmap provide a useful visualization of the relationships among the songs based on their audio features. It suggests that some features, such as tempo and pitch, are more useful than others for clustering classical music songs based on their audio features.

What are the differences in tempograms between an outlier and a song that is representative of the rest of the playlist


These four plots represent two different playlists: Iconic Soundtracks and Classical Essentials. The tempograms are a way to visualize the tempo of a piece of music over time.

The first plot shows the tempo of ‘Job, a Masque for Dancing’ by Ralph Vaughan Williams, an outlier from the Classical Essentials playlist. The second plot shows the tempo of ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’ by André Previn, a piece from the Classical Essentials playlist. Comparing the two Classical Essentials plots, we can see that ‘Job, a Masque for Dancing’ has a much more variable tempo than ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’, with many more peaks and valleys in the tempo curve. This suggests that ‘Job, a Masque for Dancing’ has a more complex rhythmic structure than ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’.

The third plot shows the tempo of ‘End Titles’ by Rachel Portman, a representative example of the rest of the Iconic Soundtracks playlist. The fourth plot shows the tempo of ‘Brooks was here’ by Thomas Newman, an outlier from the Iconic Soundtracks playlist.’End Titles’ has a relatively stable and consistent tempo throughout the entire track. On the other hand, ‘Brooks was here’, the outlier track, has a more variable tempo, with frequent changes in tempo throughout the track. Additionally, the tempo of ‘Brooks was here’ generally tends to be slower than the tempo of ‘End Titles’.

Comparing ‘End Titles’ from the Iconic Soundtracks playlist to ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’ from the Classical Essentials playlist, we can see that ‘End Titles’ has a more consistent and stable tempo than ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’, which has more fluctuations in tempo. This difference may be due to the fact that the Iconic Soundtracks playlist includes music from a wider range of genres and styles than the Classical Essentials playlist, which may have a more narrow focus on classical music.

What are the differences in keygrams between an outlier and a song that is representative of the rest of the playlist


The top two chordograms are two songs from the Iconic Soundtracks playlist, namely “Brooks was here” by Thomas Newman and “End Titles” by Rachel Portman. The chordograms represent the chords used in the songs over time, with the x-axis representing time in seconds and the y-axis representing the different chords. The chords are represented using 1-0 coding for chord templates and the Krumhansl-Kessler key profiles. The two chordograms show that the two songs have different chord progressions. “Brooks was here” has more varied and complex chords compared to “End Titles”, which has a simpler chord progression. The chordogram for “Brooks was here” has more vertical lines, indicating more chord changes, whereas the chordogram for “End Titles” has more horizontal lines, indicating longer sections with the same chord. Overall, the chordograms provide a visual representation of the harmonic structure of the two songs, highlighting differences in chord progression and complexity. It is worth noting that “Brooks was here” stands out as an outlier of the Classical Essentials playlist, with a significantly different chord progression compared to the rest of the playlist. “End Titles” serves as a representative example of the typical chord progression in the playlist.

The two chordograms on the bottom are from two songs, ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’ by André Previn, and ‘Job, a Masque for Dancing’ by Ralph Vaughan Williams. The chordograms are visual representations of the harmonic structure of the songs, with again time on the x-axis and the chords on the y-axis. The colors of the rectangles represent the strength of each chord based on the Krumhansl-Kessler key profile. The two chordograms are significantly different. The chordogram of ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’ has more consistent coloring and clear harmonic patterns. On the other hand, the chordogram of ‘Job, a Masque for Dancing’ is more complex, with a wider range of colors and less clear harmonic patterns. This difference is due to the fact that ‘Tsar Saltan, Op. 57: Flight of the Bumblebee’ is a piece from the Classical Essentials playlist and follows traditional harmonic patterns, while ‘Job, a Masque for Dancing’ is an outlier from the playlist and has more complex and unpredictable harmonic patterns.

Chroma and Timbre values of The Blade Runnner Blues


Chromagram: The distribution of the 12 pitch classes over time in the selected song.
Cepstrogram: The distribution of the 12 cepstral coefficients over time, which are related to the timbre of the sound. The combined plot allows for a comparison of the two features, highlighting the rhythmic and tonal characteristics of the song. When hovering over the chromagram plot or cepstrogram plot, the x values show the pitch class or cepstral coefficient, the y value represents the time in seconds and the z value represents the intensity of the pitch class or cepstral coefficient at the given time.

Looking at the cepstrogram plot, we can see that the song “Blade Runner Blues” has a relatively smooth and uniform distribution of timbre features over time, with a prominent peak at the lower cepstral coefficients. This suggests a relatively low level of roughness in the sound, which is consistent with the overall mellow and atmospheric mood of the song.

In terms of brightness and warmth, the cepstrogram does not reveal any particularly strong or distinctive patterns, which could suggest that these timbral features are not as salient in this song as other characteristics such as the use of ambient textures and electronic instrumentation.

Given that “Blade Runner Blues” is part of a playlist of iconic soundtracks, it is worth noting that the song was composed by Vangelis for the soundtrack of the 1982 film “Blade Runner”, which is considered a landmark of science fiction cinema. The use of electronic instruments and atmospheric textures in the song is consistent with the film’s dystopian and futuristic themes, while the melancholic and introspective mood of the music reflects the emotional depth and complexity of the film’s characters and themes.

Overall, the chromagram and cepstrogram plots provide valuable insights into the melodic and timbral characteristics of “Blade Runner Blues”, which can be used to shed light on the song’s structure, style, and meaning, as well as its cultural and historical significance as part of the iconic “Blade Runner” soundtrack.

Chroma Values Soundstrack outlier


These plots show the chroma features for ‘Blade Runner Blues’ from the “iconic_soundtracks” Spotify playlist and …

The chroma feature is a way of representing the tonal content of an audio signal. It is based on the 12 different pitches in a chromatic scale (C, C#, D, D#, E, F, F#, G, G#, A, A#, B), and for each pitch it calculates a value that represents the amount of energy in the audio signal that corresponds to that pitch.

In this plot, the x-axis shows the 12 different pitches (notes) in the chromatic scale, and the y-axis shows the corresponding chroma value for each note. The height of each bar represents the amount of energy in the audio signal that corresponds to that note, and the color of each bar indicates the magnitude of the chroma value (with darker colors indicating higher values). The hover text displays the exact percentage of energy in the audio signal that corresponds to each note.

In summary, this plot shows the relative distribution of energy across the 12 different pitches in the audio signal for a specific track.

Chroma Values Classical outlier


These plots show the chroma features for ‘Blade Runner Blues’ from the “iconic_soundtracks” Spotify playlist and …

The chroma feature is a way of representing the tonal content of an audio signal. It is based on the 12 different pitches in a chromatic scale (C, C#, D, D#, E, F, F#, G, G#, A, A#, B), and for each pitch it calculates a value that represents the amount of energy in the audio signal that corresponds to that pitch.

In this plot, the x-axis shows the 12 different pitches (notes) in the chromatic scale, and the y-axis shows the corresponding chroma value for each note. The height of each bar represents the amount of energy in the audio signal that corresponds to that note, and the color of each bar indicates the magnitude of the chroma value (with darker colors indicating higher values). The hover text displays the exact percentage of energy in the audio signal that corresponds to each note.

In summary, this plot shows the relative distribution of energy across the 12 different pitches in the audio signal for a specific track.